ouroboros/devtools/benchmarks/common
Anton Razzhigaev db910f5fc3 Harden CyberGym isolated settings and archive publication
Bind applied settings to the isolated child, keep the snapshot immutable, and make archive publication and cleanup descriptor-safe across replacement races. Add focused regression coverage and document the three-task smoke contract.

Co-authored-by: Ouroboros <311266734+ouroboros-agent@users.noreply.github.com>
2026-08-28 04:49:41 +00:00
..
__init__.py feat(devtools-benchmarks): add official benchmark harnesses and workspace executor 2026-06-06 12:03:30 +03:00
launcher_audit.py Add CyberGym benchmark profile and methodology 2026-08-26 23:44:07 +00:00
manifests.py Harden CyberGym settings provenance and isolation 2026-08-28 03:01:00 +00:00
model_slots.py Bind Harbor smoke to its effective Light actor 2026-08-20 04:52:26 +03:00
official_commands.py feat: v6.55.0 bench devtools alignment — scaffold defaults, ProgramBench e2e, CLB launcher, OSWorld 2.0 2026-07-03 22:49:01 +03:00
result_index.py Review-fix batch (Sol delta review on 50e824fe..d18d7ebf): owner-stop deadline honesty 2026-08-15 19:19:57 +03:00
run_roots.py feat(v6.81.0): benchmark provenance becomes a gate, and the artefacts stop lying 2026-07-26 03:40:44 +00:00
secrets.py Harden CyberGym single-model settings provenance 2026-08-28 02:45:02 +00:00
server_runner.py Harden CyberGym isolated settings and archive publication 2026-08-28 04:49:41 +00:00
subprocesses.py feat(devtools-benchmarks): add official benchmark harnesses and workspace executor 2026-06-06 12:03:30 +03:00